The COLUMN list is used with the sparse input format of side constraints. This list consists of one SAS variable in the CONDATA= data set that has as values the names of arc variables, nonarc variables, or missing values. Some, if not all of these values, also can be values of the NAME list variables of the ARCDATA= data set. The COLUMN list variable can have other special values (refer to the TYPEOBS= and RHSOBS= options). If the COLUMN list is not specified after the PROC NETFLOW statement, the CONDATA= data set is searched and a SAS variable named _COLUMN_ is used. The COLUMN list variable must have character values.